U.S. patent application number 16/179807 was filed with the patent office on 2019-05-09 for system and method for breathing pattern extraction from ppg signals.
This patent application is currently assigned to Tata Consultancy Services Limited. The applicant listed for this patent is Tata Consultancy Services Limited. Invention is credited to Avik GHOSE, Dibyanshu JAISWAL, Dhaval Satish JANI, Shalini MUKHOPADHYAY.
Application Number | 20190133537 16/179807 |
Document ID | / |
Family ID | 64401966 |
Filed Date | 2019-05-09 |
![](/patent/app/20190133537/US20190133537A1-20190509-D00000.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00001.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00002.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00003.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00004.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00005.png)
![](/patent/app/20190133537/US20190133537A1-20190509-D00006.png)
![](/patent/app/20190133537/US20190133537A1-20190509-M00001.png)
United States Patent
Application |
20190133537 |
Kind Code |
A1 |
GHOSE; Avik ; et
al. |
May 9, 2019 |
SYSTEM AND METHOD FOR BREATHING PATTERN EXTRACTION FROM PPG
SIGNALS
Abstract
A system and method for extracting breathing patterns from PPG
signals are provided. The method includes designing a filter for
extracting breathing patterns from PPG signals. Designing the
filter includes defining filter specifications for extraction of
breathing pattern from the PPG signals. Herein, the filter
specifications includes a type, an order and a cut-off frequency of
the filter. Designing the filter further includes generating a
transfer function associated with the filter specifications, and
computing a plurality of filter coefficients using filtfilt
function for allowing filtering of the PPG signals. Using the
filter comprising the plurality of filter coefficients, a filtered
PPG signal is generated by removing DC component from PPG signals
obtained from a wearable device being worn by a subject. The
filtered PPG signal is indicative of the breathing pattern of the
subject.
Inventors: |
GHOSE; Avik; (Kolkata,
IN) ; MUKHOPADHYAY; Shalini; (Kolkata, IN) ;
JAISWAL; Dibyanshu; (Kolkata, IN) ; JANI; Dhaval
Satish; (Rockville, MD) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Tata Consultancy Services Limited |
Mumbai |
|
IN |
|
|
Assignee: |
Tata Consultancy Services
Limited
Mumbai
IN
|
Family ID: |
64401966 |
Appl. No.: |
16/179807 |
Filed: |
November 2, 2018 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
A61B 5/742 20130101;
A61B 5/7207 20130101; A61B 5/7278 20130101; A61B 5/721 20130101;
A61B 5/7257 20130101; A61B 5/681 20130101; A61B 5/02438 20130101;
A61B 5/7221 20130101; A61B 5/0816 20130101; A61B 5/02416 20130101;
A61B 2562/0219 20130101; A61B 5/7203 20130101; A61B 5/725
20130101 |
International
Class: |
A61B 5/00 20060101
A61B005/00; A61B 5/08 20060101 A61B005/08; A61B 5/024 20060101
A61B005/024 |
Foreign Application Data
Date |
Code |
Application Number |
Nov 3, 2017 |
IN |
201721039317 |
Claims
1. A processor-implemented method for extracting breathing patterns
from PPG signals, the method comprising: designing, via one or more
hardware processors, a filter for extracting breathing patterns
from PPG signals, wherein designing the filter comprises: defining
filter specifications for extraction of breathing pattern from the
PPG signals, wherein the filter specifications comprises a type, an
order and a cut-off frequency of the filter, generating a transfer
function associated with the filter specifications, and computing a
plurality of filter coefficients using filtfilt function for
allowing filtering of the PPG signals; generating, using the filter
comprising the plurality of filter coefficients, filtered PPG
signals by removing DC component from PPG signals obtained from a
wearable device worn by a subject via the one or more hardware
processors, the filtered PPG signals indicative of the breathing
pattern of the subject.
2. The method as claimed in claim 1, further comprising: obtaining,
in real time, raw PPG signals from the wearable device worn by the
subject; and pre-processing the raw PPG signals to obtain the PPG
signals associated with a window size and a sampling frequency
suitable for the breathing pattern extraction.
3. The method as claimed in claim 2, wherein the window size is
around 10 seconds.
4. The method as claimed in claim 1, wherein the filter
specifications for the extraction of the breathing pattern
comprises Infinite Impulse response (IIR) Butter-worth band-pass
filter of order 4, with a cut-off frequency of 0.2-1 Hz.
5. The method as claimed in claim 1, further comprising: obtaining,
from an accelerometer sensor embodied in the wearable device, an
accelerometer signal indicative of motion of the subject; computing
mean and standard deviation values of the accelerometer signal to
determine whether the subject is in motion; and performing, based
on the determination, one of: discarding the PPG signals captured
from the wearable device upon determination of the subject to be in
motion, and generating the plurality of filter coefficients upon
determination of the subject to be in rest position
6. A system for extracting breathing patterns from PPG signals, the
system comprising: one or more memories; and one or more hardware
processors, the one or more memories coupled to the one or more
hardware processors, wherein the one or more hardware processors
are capable of executing programmed instructions stored in the one
or more memories to: design a filter for extracting breathing
patterns from the PPG signals, wherein designing the filter
comprises: defining filter specifications for extraction of
breathing pattern from the PPG signals, wherein the filter
specifications comprising a type, an order and a cut-off frequency
of the filter, generating a transfer function associated with the
filter specifications, and computing a plurality of filter
coefficients using filtfilt function for allowing filtering of the
PPG signals; generate, using the filter comprising the plurality of
filter coefficients, filtered PPG signals by removing DC component
from the PPG signals obtained from a wearable device worn by a
subject, the filtered PPG signals indicative of the breathing
pattern of the subject.
7. The system as claimed in claim 6, wherein one or more hardware
processors are further configured by the instructions to: obtain,
in real time, raw PPG signals from the wearable device being worn
by the subject; pre-process the raw PPG signals to obtain the PPG
signals associated with a window size and a sampling frequency
suitable for the breathing pattern extraction.
8. The system as claimed in claim 7, wherein the window size is
around 10 seconds.
9. The system as claimed in claim 6, wherein the filter
specifications for the extraction of the breathing pattern
comprises Infinite Impulse Response (IIR) Butter-worth band-pass
filter of order 4, with a cut-off frequency of 0.2-1 Hz.
10. The system as claimed in claim 6, wherein the one or more
hardware processors are further configured by the instructions to:
obtain, from an accelerometer sensor embodied in the wearable
device, an accelerometer signal indicative of motion of the
subject; compute mean and standard deviation values of he
accelerometer signal to determine whether the subject is in motion;
and perform, based on the determination, one of: discard the PPG
signals captured from the wearable device upon determination of the
subject to be in motion, and generate the plurality of filter
coefficients upon determination of the subject to be in rest
position.
11. One or more non-transitory machine readable information storage
mediums comprising one or more instructions which when executed by
one or more hardware processors causes the one or more hardware
processor to perform a method for extracting breathing patterns
from PPG signals, said method comprising: designing, via the one or
more hardware processors, a filter for extracting breathing
patterns from PPG signals, wherein designing the filter comprises:
defining filter specifications for extraction of breathing pattern
from the PPG signals, wherein the filter specifications comprises a
type, an order and a cut-off frequency of the filter, generating a
transfer function associated with the filter specifications, and
computing a plurality of filter coefficients using filtfilt
function for allowing filtering of the PPG signals; generating,
using the filter comprising the plurality of filter coefficients,
filtered PPG signals by removing DC component from PPG signals
obtained from a wearable device worn by a subject via the one or
more hardware processors, the filtered PPG signals indicative of
the breathing pattern of the subject.
12. The one or more non-transitory machine readable information
storage mediums of claim 11, further comprising: obtaining, in real
lime, raw PPG signals from the wearable device worn by the subject;
and pre-processing the raw PPG signals to obtain the PPG signals
associated with a window size and a sampling frequency suitable for
the breathing pattern extraction.
13. The one or more non-transitory machine readable information
storage mediums of claim 12, wherein the window size is around 10
seconds.
14. The one or more non-transitory machine readable information
storage mediums of claim 11 wherein the filter specifications for
the extraction of the breathing pattern comprises Infinite Impulse
response (IIR) Butter-worth band-pass filter of order 4, with a
cut-off frequency of 0.2-1 Hz.
15. The one or more non-transitory machine readable information
storage mediums of claim 11, further comprising: obtaining, from an
accelerometer sensor embodied in the wearable device, an
accelerometer signal indicative of motion of the subject; computing
mean and standard deviation values of the accelerometer signal to
determine whether the subject is in motion; and performing, based
on the determination, one of: discarding the PPG signals captured
from the wearable device upon determination of the subject to be in
motion, and generating the plurality of filter coefficients upon
determination of the subject to be in rest position
Description
PRIORITY CLAIM
[0001] This U.S. patent application claims priority under 35 U.S.C.
.sctn. 119 to: India Application No. 201721039317, filed on Nov. 3,
2017. The entire contents of the aforementioned application are
incorporated herein by reference.
TECHNICAL FIELD
[0002] The present disclosure in general relates to extraction of
breathing patterns from photoplethysmogram (PPG) signals, and more
particularly to system and method for designing a filter to extract
breathing patterns from PPG signals.
BACKGROUND
[0003] Today's sedentary work environment and an unhealthy eating
lifestyle has attracted a spectrum of cardiopulmonary diseases. As
of 2017, cardiac diseases are responsible for maximum deaths in
United States, with chronic respiratory disorders ranking no. 3.
According to the CDC, more than 40% of such disorders lead to
deaths outside the hospital, the reason being failure to detect
early warning signs. It has thus become imperative to be able to
not only detect the actual symptoms of such disorders beforehand,
but also to detect the possibility of any abnormality in the
cardiopulmonary system of the body.
[0004] Cardiopulmonary Exercise Test (CPET) is an important
clinical tool in detecting cardiac stress levels in subjects to be
tested for heart and lung disease, or the patients scheduled for a
major surgery. CPDET is an involved test which requires the patient
to breathe into a special mouthpiece and Electro cardiogram (ECG)
of the subject is recorded before, during and after a stationary
workout, e.g., on-the-spot cycling. Such tests can provide accurate
and detailed heart and lung performance of the subject.
[0005] The inventors here have recognized several technical
problems with such conventional tests, as explained below. Due to
the use of specialised equipment, such test pose practical
restrictions on their usage. People may fail to take such
specialised tests at the onset of concerned disorders, which may
worsen with time. Moreover, Current solutions pertaining to PPG
based breathing monitoring are primarily able to compute breathing
rate. However, there may be more markers to cardiac and lung
problems hidden in the details of a breathing cycle. There exists
sensors like Tidal Breathing Pattern Recorder (TBPR) however, these
are invasive tests where the subject needs to blow into a pipe.
Moreover, ubiquitous round the clock monitoring of breathing cycles
is not available in state of art.
SUMMARY
[0006] The following presents a simplified summary of some
embodiments of the disclosure in order to provide a basic
understanding of the embodiments. This summary is not an extensive
overview of the embodiments. It is not intended to identify
key/critical elements of the embodiments or to delineate the scope
of the embodiments. Its sole purpose is to present some embodiments
in a simplified form as a prelude to the more detailed description
that is presented below.
[0007] In view of the foregoing, an embodiment herein provides
method and system for extracting breathing patterns from PPG
signals. The method includes designing, via one or more hardware
processors, a filter for extracting breathing patterns from PPG
signals. Herein designing the filter includes defining filter
specifications for extraction of breathing pattern from the PPG
signals, wherein the filter specifications comprising a type, an
order and a cut-off frequency of the filter. Further designing the
filter includes generating a transfer function associated with the
filter specifications, and [0008] computing a plurality of filter
coefficients using filtfilt function for allowing filtering of the
PPG signals. The method further includes generating, using the
filter comprising the plurality of filter coefficients, a filtered
PPG signal by removing DC component from PPG signals obtained from
a wearable device being worn by a subject via the one or more
hardware processors. The filtered PPG signal indicative of the
breathing pattern of the subject.
[0009] In another aspect, a system for extracting breathing
patterns from PPG signals is provided. The system includes one or
more memories; and one or more hardware processors, the one or more
memories coupled to the one or more hardware processors, wherein
the one or more hardware processors are capable of executing
programmed instructions stored in the one or more memories to
design a filter for extracting breathing patterns from PPG signals.
To design the filter, the one or more hardware processors are
configured by the instructions to define filter specifications for
extraction of breathing pattern from the PPG signals, wherein the
filter specifications comprising a type, an order and a cut-off
frequency of the filter, generate a transfer function associated
with the filter specifications, and compute a plurality of filter
coefficients using filtfilt function for allowing filtering of the
PPG signals. Further, the one or more hardware processors are
further configured by the instructions to generate, using the
filter comprising the plurality of filter coefficients, a filtered
PPG signal by removing DC component from PPG signals obtained from
a wearable device being worn by a subject. The filtered PPG signals
are indicative of the breathing pattern of the subject.
[0010] In yet another aspect, a non-transitory computer-readable
medium having embodied thereon a computer program for executing a
method for extracting breathing patterns from PPG signals. The
method includes designing, via one or more hardware processors, a
filter for extracting breathing patterns from PPG signals. Herein
designing the filter includes defining filter specifications for
extraction of breathing pattern from the PPG signals, wherein the
filter specifications comprising a type, an order and a cut-off
frequency of the filter. Further designing the filter includes
generating a transfer function associated with the filter
specifications, and computing a plurality of filter coefficients
using filtfilt function for allowing filtering of the PPG signals.
The method further includes generating, using the filter comprising
the plurality of filter coefficients, a filtered PPG signal by
removing DC component from PPG signals obtained from a wearable
device being worn by a subject via the one or more hardware
processors. The filtered PPG signal indicative of the breathing
pattern of the subject.
BRIEF DESCRIPTION OF THE FIGURES
[0011] The detailed description is described with reference to the
accompanying figures. In the figures, the left-most digit(s) of a
reference number identifies the figure in which the reference
number first appears. The same numbers are used throughout the
drawings to reference like features and modules.
[0012] FIG. 1A illustrates a networking environment implementing
system for extraction of breathing patterns from PPG signals, in
accordance with an example embodiment.
[0013] FIG. 1B illustrates a PPG Signal and its Corresponding
Dual-Peaked Frequency Equivalent, in accordance with an example
embodiment.
[0014] FIG. 2 illustrates a block diagram of a system for
extraction of breathing patterns from PPG signals, in accordance
with an example embodiment.
[0015] FIG. 3 illustrate a Coherence in breathing patterns between
that extracted from PPG and TBPS, in accordance with an example
embodiment.
[0016] FIG. 4 illustrates an example flow diagram for a method for
extraction of breathing patterns from PPG signals, in accordance
with an example embodiment.
[0017] FIG. 5 illustrates an example flow diagram for a method for
extraction of breathing patterns from PPG signals, in accordance
with an example embodiment.
[0018] FIG. 6 illustrates PC calculated for each breathing cycle
for a single best session, in accordance with an example
embodiment.
[0019] FIG. 7 illustrates PC calculated for each full session, in
accordance with an example embodiment.
[0020] FIG. 8 illustrates mean absolute deviation in breathing rate
for all sessions, in accordance with an example embodiment.
[0021] It should be appreciated by those skilled in the art that
any block diagrams herein represent conceptual views of
illustrative systems and devices embodying the principles of the
present subject matter. Similarly, it will be appreciated that any
flow charts, flow diagrams, and the like represent various
processes which may be substantially represented in computer
readable medium and so executed by a computer or processor, whether
or not such computer or processor is explicitly shown.
DETAILED DESCRIPTION
[0022] Cardiopulmonary disease prognosis can achieve therapeutic
edge if the disorders can be detected and attended to at an early
stage. Currently, Cardiopulmonary Exercise Test (CPET) acts as an
important clinical tool in detecting cardiac stress levels in
subjects to be tested for heart and lung diseases, or the patients
scheduled for a major surgery. However, people may fail to take
such specialised tests at the onset of concerned disorders, which
may worsen with time.
[0023] According to American College of Cardiology
Foundation/American Heart Association (ACCF/AHA), the most
noteworthy and common symptoms of any cardiac disorder is shortness
of breath (dyspnea) and fatigue coupled with palpitation if the
body is subjected to an above-ordinary exertion. If the major
physiological parameters can be recorded and studied during such
spells while the subject is unobtrusively involved in routine
activities, the need for specialised tests as discussed previously,
can be eliminated and any anomalous behaviour in the stress-bearing
capability of the body can be detected, hence providing the signs
for any possible cardiopulmonary disorder at an earliest stage may
be possible.
[0024] New York Heart Association (NYHA) classifies subjects into
different classes denoting the stages in a possible cardiac
disorder, where a Functional Capacity II, Objective Assessment B
refers to the patients exhibiting dyspnoea, palpitation and fatigue
in physical activities, which is the target detection class of the
proposed platform. Said symptoms are almost ubiquitous among the
cardiopulmonary disorders which are also reflected in the ACCF/AHA
staging system where level C includes all the NYHA classes, and
presents fatigue and dyspnoea as associated symptoms.
[0025] Estimation of physiological parameters like heart rate (HR)
and blood pressure (BP) using PPG from mobile devices have been of
prime interest to the researches. Using dedicated devices such as a
pulse oximeter pose practical restrictions on automatic and
on-the-go monitoring preventing their adoption in everyday life.
Additionally, mobile phone camera has been extensively used in
order to extract PPG signal and process it for heart rate
estimation. There have been many researches targeting wearable
device platform for extraction of PPG signals. In a conventional
method, a reflective technique is used for extraction of PPG
signals by using an ear-worn sensor to address application sensor
variations. Another conventional method discloses the concept of
wrist-worn wearables (or wearable devices) that have been
researched extensively for HR and BP estimation.
[0026] In recent years, powerful devices such as smartwatches
equipped with dedicated PPG sensors have been introduced. Said
devices have quickly emerged as a choice of stylish wearables among
the users, which have encouraged the research community to extend
the PPG based research in this direction.
[0027] One conventional system uses FFT technique to infer
breathing rate from the PPG cycles. Another conventional approach
has tried to reproduce breathing cycles from PPG signals using
MSPCA which is a wavelet-based adaptive multiscale principal
component analysis, however such technique is not computationally
very viable. Said conventional techniques and/or systems however
aims at extracting respiration cycles from PPG with computational
efficiency by enabling it to run on relatively constrained
environments of smartwatches in real-time.
[0028] Various embodiments disclosed herein provides method and
system for round-the-clock monitoring of a person's breathing
pattern, and provides them with the breathing rate values in an
unobtrusive manner, utilizing the PPG signal obtained from a
wearable device such as a wrist wearable. The disclosed method is a
computationally efficient solution that is devised to reconstruct
breathing cycles from raw PPG signal recorded from the smartwatch,
which enables the system to continuously monitor the subject's
heart rate and breathing pattern online. The method achieves high
accuracy when verified against a pre-calibrated flow-meter. As a
sensing hardware, in an exemplary embodiment, the disclosed system
utilizes a wearable device for instance a smartwatch, and at the
same time boasts a powerful set of sensors, including Inertial
Measurement Unit (IMU) and photoplethysmogram (PPG) sensor. Hence,
the user is not subject to any special hardware. The system
utilizes IMU in order to detect the spells of intensive activities,
and uses PPG to extract the cardiac and breathing cycle
information.
[0029] An important contribution of the disclosed method and system
is to effectively design a filter for extraction of breathing
patterns. A detailed description of the above described system is
shown with respect to illustrations represented with reference to
FIGS. 1A through 8.
[0030] The method(s) and system(s) for quality extraction of
breathing patterns are further described in conjunction with the
following figures. It should be noted that the description and
figures merely illustrate the principles of the present subject
matter. It will thus be appreciated that those skilled in the art
will be able to devise various arrangements that, although not
explicitly described or shown herein, embody the principles of the
present subject matter and are included within its spirit and
scope. Furthermore, all examples recited herein are principally
intended expressly to be only for pedagogical purposes to aid the
reader in understanding the principles of the present subject
matter and the concepts contributed by the inventor(s) to
furthering the art, and are to be construed as being without
limitation to such specifically recited examples and conditions.
Moreover, all statements herein reciting principles, aspects, and
embodiments of the present subject matter, as well as specific
examples thereof, are intended to encompass equivalents
thereof.
[0031] FIG. 1A illustrates a network environment 100 implementing a
system 102 for extraction of breathing patterns using PPG signals,
according to an embodiment of the present subject matter. It will
be noted herein that the PPG signals are considered herein for the
extraction of breathing patterns since PPG signals can be very
effective in determining cardiac cycle due to their
non-invasiveness and accurate results. PPG signals also contain
information about the systolic/diastolic pressure difference in
alternating current (AC) component thereof. It is also possible to
extract the respiratory cycle from PPG. The PPG sensor embodied in
a wearable device such as a smartwatch facilitates in physiological
monitoring in a non-obtrusive manner, since it does not require
action from user's perspective. Thus, the disclosed embodiments are
capable of extracting breathing patterns from the PPG signals
derived from user's wearable devices such as smartwatch (when worn
by the user), as will be explained further in detail below.
[0032] In an embodiment, the system 102 may receive inputs from the
user's wearable device in form of PPG signals. There is a high
coherence between the respiratory cycles and PPG signals. A typical
PPG signal and its corresponding frequency spectrum is shown in
FIG. 1B, where the peaks pertaining to heart rate and breathing
rate can be seen. In order to enable a detailed analysis at a later
stage, full breathing cycles can prove to be more useful than only
the respiratory rate. In an embodiment, the system 102 for
monitoring may be embodied on the wearable device such as the
smartwatch, without any special add-on hardware requirement,
thereby providing a computationally efficient way of extracting
breathing patterns from the PPG signals.
[0033] In an embodiment, raw PPG data is collected from the PPG
sensor embodied in the wearable device (such as the smartwatch) and
is resampled to a rate of 50 Hz to obtain a resampled PPG data. The
resampled PPG data is analysed by the system on a windowed basis. A
zero phase forward-reverse filtering is performed on the resampled
PPG data using a Butterworth HR filter. Typically, a person's
normal breathing rate falls between 15-25 breaths/min. During
fatigued spells, the shortness of breath can cause the breathing
rate to rise to 50-60 breaths/min. Hence, said filter's pass-band
is specified as 0.25 Hz -1 Hz. FIG. 3 shows a comparison between an
extracted breathing pattern from the PPG and the same from a TBPS.
It is evident that the patterns match appreciably, proving the
effectiveness of the disclosed method.
[0034] As described above, the system may be embodied in a wearable
device worn by the user. In an alternative embodiment, the system
102 may be embodied in a computing device, for instance a computing
device 104 communicably coupled with the wearable device. In an
embodiment, the computing device may be an example of a server.
Herein, although the present disclosure is explained considering
that the system 102 is implemented on a server, it may be
understood that the system 102 may also be implemented in a variety
of computing systems, such as a laptop computer, a desktop
computer, a notebook, a workstation, a cloud-based computing
environment and the like. In one implementation, the system 102 may
be implemented in a cloud-based environment. It will be understood
that the system 102 may be accessed by multiple users through one
or more user devices 106-1, 106-2 . . . 106-N, collectively
referred to as user devices 106 hereinafter, or applications
residing on the user devices 106. Examples of the user devices 106
may include, but are not limited to, a portable computer, a
personal digital assistant, a handheld device, a Smartphone, a
Tablet Computer, a workstation and the like. The user devices 106
are communicatively coupled to the system 102 through a network
108. Herein, the users of the user-devices 106 may include users
wearing similar wearable devices.
[0035] In an embodiment, the network 108 may be a wireless or a
wired network, or a combination thereof. In an example, the network
108 can be implemented as a computer network, as one of the
different types of networks, such as virtual private network (VPN),
intranet, local area network (LAN), wide area network (WAN), the
internet, and such. The network 106 may either be a dedicated
network or a shared network, which represents an association of the
different types of networks that use a variety of protocols, for
example, Hypertext Transfer Protocol (HTTP), Transmission Control
Protocol/Internet Protocol (TCP/IP). and Wireless Application
Protocol (WAP), to communicate with each other. Further, the
network 108 may include a variety of network devices, including
routers, bridges, servers, computing devices, storage devices. The
network devices within the network 108 may interact with the system
102 through communication links.
[0036] As discussed above, the system 102 may be implemented in a
computing device 104, such as a hand-held device, a laptop or other
portable computer, a tablet computer, a mobile phone, a FDA, a
smartphone, and a wearable device such as a smart watch. The system
102 may also be implemented in a workstation, a mainframe computer,
a server, and a network server. In an embodiment, the system 102
may be coupled to a data repository, for example, a repository 112.
The repository 112 may store data processed, received, and
generated by the system 102. In an alternate embodiment, the system
102 may include the data repository 112. The components and
functionalities of the system 102 are described further in detail
with reference to FIG. 2.
[0037] FIG. 2 illustrates a block diagram of a system 200 for
breathing pattern extraction from PPG signals, in accordance with
an example embodiment. The system 200 may be an example of the
system 102 (FIG. 1A). In an example embodiment, the system 200 may
be embodied in, or is in direct communication with the system, for
example the system 102 (FIG. 1). It will be understood that the
system 200 for breathing pattern extraction can be used in various
applications, including but not limited to, intensive activity
detection, fatigue detection, and so on.
[0038] The system 200 includes or is otherwise in communication
with one or more hardware processors such as a processor 202, at
least one memory such as a memory 204, and an I/O interface 206.
The processor 202, memory 204, and the I/O interface 206 may be
coupled by a system bus such as a system bus 208 or a similar
mechanism. The I/O interface 206 may include a variety of software
and hardware interfaces, for example, a web interface, a graphical
user interface, and the like The interfaces 206 may include a
variety of software and hardware interfaces, for example,
interfaces for peripheral device(s), such as a keyboard, a mouse,
an external memory, a camera device, and a printer. Further, the
interfaces 206 may enable the system 102 to communicate with other
devices, such as web servers and external databases. The interfaces
206 can facilitate multiple communications within a wide variety of
networks and protocol types, including wired networks, for example,
local area network (LAN), cable, etc., and wireless networks, such
as Wireless LAN (WLAN), cellular, or satellite. For the purpose,
the interfaces 206 may include one or more ports for connecting a
number of computing systems with one another or to another server
computer. The I/O interface 206 may include one or more ports for
connecting a number of devices to one another or to another
server.
[0039] The hardware processor 202 may be implemented as one or more
microprocessors, microcomputers, microcontrollers, digital signal
processors, central processing units, state machines, logic
circuitries, and/or any devices that manipulate signals based on
operational instructions. Among other capabilities, the hardware
processor 202 is configured to fetch and execute computer-readable
instructions stored in the memory 204.
[0040] The memory 204 may include any computer-readable medium
known in the art including, for example, volatile memory, such as
static random access memory (SRAM) and dynamic random access memory
(DRAM), and/or non-volatile memory, such as read only memory (ROM),
erasable programmable ROM, flash memories, hard disks, optical
disks, and magnetic tapes. In an embodiment, the memory 204
includes a plurality of modules 220 and a repository 240 for
storing data processed, received, and generated by one or more of
the modules 220. The modules 220 may include routines, programs,
objects, components, data structures, and so on, which perform
particular tasks or implement particular abstract data types.
[0041] The repository 240, amongst other things, includes a system
database 242 and other data 244. The other data 244 may include
data generated as a result of the execution of one or more modules
in the modules 220. The repository 240 is further configured to
maintain PPG data 246 and motion data 248. The PPG data 246 may be
obtained from PPG sensor(s) 252 embodied in the wearable device,
such as wearable device 250 and may contain data associated with
raw PPG signal. The motion data 248 may contain data associated
with intense activity being performed by the user. The motion data
248 may be captured from sensors such as accelerometer 254 and
barometer embodied in the wearable device 250. The motion data
provide information associated with the intense activity being
performed by the user. Herein, it will be understood that if the
major physiological parameters associated with a user can be
recorded and studied while the subject is unobtrusively involved in
activities, the need for specialised tests to detect abnormal
stress causing condition, can be eliminated and any anomalous
behavior in the stress-bearing capability of the body can be
detected, hence provides the signs for any possible cardiopulmonary
disorder at an earliest stage possible. For example, such
monitoring of parameters can disclose/detect the patients
exhibiting dyspnea, palpitation and fatigue in physical activities,
and so on.
[0042] In an embodiment, the system 200 takes raw PPG signals (or
PPG data) and motion data obtained from user's wearable device as
input for pre-processing. The PPG signal taken from the wrist
wearable may contain noise and may be distorted, and accordingly
may be difficult to work with. Hence, the interpolated and windowed
signal is then pre-processed, i.e., DC component of the signal is
removed, and the signal is filtered by a filter.
[0043] Pre-processing the raw PPG signals provides PPG signals
associated with a window size and a sampling frequency suitable for
the breathing pattern extraction. In an embodiment, the
pre-processing the raw PPG signal and the motion data includes
interpolating signals received from the PPG sensor and motion
sensors to bring them to a uniform sampling frequency, F.sub.s of
50 Hz. The interpolated PPG signals are then taken window-wise, for
further analysis, as explained further below in the description.
Herein, the window size or windows of interest refers to the
windows in which it is preferred to analyse PPG data so as to
compute physiology of the subject. In general, such windows of
interest correspond to the subject being in rest position before
and after performing some activity. As explained above, to detect a
subject is engaged in an activity, the system 200 receives inputs
from the motion sensors such as accelerometer and/or barometer.
Accelerometer is a sensor that can detect any acceleration force
exerted on the (3-axes: x, y, z of the) device. To detect the state
of the device (i.e. whether the device worn by the user is in rest
or in motion), the system 200 calculates the resultant value of the
accelerometer readings. In the rest position of the user, the
resultant value is nearly equal to g (i.e. 9.8 m/s.sup.2) whereas
during motion the resultant value fluctuates between 0 to +4 g.
Since the wearable device is worn by the subject, by the help of
accelerometer sensor, the system can distinctively detect motion
and rest phases of the subject. In an embodiment, the system 200
computes mean and standard deviation of the accelerometer signal to
determine whether the subject is in motion or at rest. If the
subject is determined to be in motion, the PPG signal captured
corresponding to said window is discarded. If, however, the system
determines the subject to be at rest or in rest position, the PPG
signal captured corresponding to said window is considered for
further processing for breathing pattern extraction. In an
embodiment, said further processing includes generating a plurality
of filter coefficients, as will be described later in the
description.
[0044] In an example embodiment, the window size may be around 10
seconds. In an embodiment, since the wearable device is
continuously worn by the user, the system 200 is able to
continuously monitor the inputs such as the PPG signal of the user,
for example, during rest and during motion. Elongated periods of
rest (such as 10 minutes or more), are one type of windows of
interest, whereby the system can detect physiological parameters of
the user when the user is normal and relaxed. The other types of
windows of interest may be just after the user has performed some
activity for some time. The reason for selection of said period as
window of interest is because, there are visible changes in the
physiological parameters (increased Heart rate and breathing rate)
of the user just after performing some strenuous activity (like
climbing stairs, running, and so on). The changes observed in the
physiological parameters with respect to rest and just after motion
windows of interest can be possibly used to study for non-specific
bio-markers.
[0045] Herein, the choice of the window size is an important
factor, as it determines the accuracy of the system but in the
present case, there is a trade-off between accuracy, latency and
number of potent measurements. For example, if window size is say
30 sec, the accuracy is +-1 BrPM, but due to large window size the
system has high latency i.e. the system needs to wait for long (30
secs) to obtain one output. Hence, it is important to decide not to
sacrifice on accuracy to get quick results or get high accuracy for
late results. Since all the accuracy of +-3 BrPM is acceptable, the
system can be designed 10 seconds window size. Here, 10 seconds
window size is considered to attain an accuracy of +-3 BrPM, and
get the BR values at a fairly acceptable interval, thereby solving
the accuracy versus latency trade-off for a healthcare
application
[0046] As previously described, upon determination of the subject
to be at rest, the system generates the plurality of filter
coefficients for filtering the processed PPG signal. In an
embodiment, the system generates said filter coefficients by
removing DC component from the PPG signals. Herein, the filtered
PPG signal indicative of the breathing pattern of the subject.
[0047] In an embodiment, the system 200 is capable of designing a
filter for extracting breathing patterns from PPG signals. Herein,
the filter is designed such that it can effectively separate out
breathing signal from a pulse-measurement sensor signal like PPG.
In an embodiment, designing the filter includes defining filter
specifications for extraction of breathing pattern from the PPG
signals. In an embodiment, the filter specifications include a
type, an order and a cut-off frequency of the filter. In an example
embodiment, the filter specifications for breathing pattern
extraction may include IR Butter-worth band-pass filter of order 4,
with a cut-off frequency of 0.2-1 Hz, as will be explained further
in detail below.
[0048] Designing a filter consists of developing specifications
appropriate to the problem (for example, a second-order low pass
filter with a specific cut-off frequency), and then producing a
transfer function which meets the specifications:
H(z)=B(z)/A(z)=(b0+b1*z -1+b2*z -2+ . . . +bN*z{circle around (
)}-N)/(1+a1*z -1+a2*z -2+ . . . +aN*z -N) (1) [0049] where, a is a
matrix 3.times.3 [a0, a1, . . . a8], and [0050] b is a matrix
3.times.3 [b1, b1, . . . b8]. [0051] Here, a and b are coefficients
of the transfer function fed as input to the filter.
[0052] This is the form for a recursive filter, which typically
leads to an infinite impulse response (IIR) behaviour, but if the
denominator is made equal to unity i.e. no feedback, then this
becomes an FIR or finite impulse response filter. In order to
achieve good results in extraction of breathing pattern and
calculation of breathing rate from a wearable-PPG signal, the
filter should be able to efficiently eliminate noise, while keeping
the computation as simple as possible. This is very important,
especially in case of breathing pattern extraction from wrist PPG
signal, as the signal inherently contains a lot of noise owing to
the wrist movements and change of position of sensor while the
subject is wearing the wrist wearable.
[0053] The important parameters that determine the quality of
filter to get the desired results are the type of filter and the
order of filter. The order of a filter usually refers to the number
of components (capacitors and inductors, not resistors or
transistors) or the number of computations required for the filter,
that affect the `steepness` or `shape` of the filter's frequency
response. A first-order recursive filter will only have a single
frequency-dependent component. This means that the slope of the
frequency response is limited to 6 dB per octave. For many
purposes, this is not sufficient. To achieve steeper slopes,
higher-order filters are required. Hence, the disclosed embodiments
have implemented the HR Butterworth Filter of order 4, which helps
in achieving a better shape of frequency response, while keeping in
mind the stability of the filter and reducing computational
complexities at the same time.
[0054] For breathing pattern extraction, the filter design includes
an IIR Butterworth band-pass filter of order 4, with a cut-off
frequency of 0.2-1 Hz, so as to include slow (12-15 BrPM) as well
as fast (50-55 BrPM) breathing rates. The primary advantage of HR
filters over FIR filters is that they typically meet a given set of
specifications with a much lower filter order than a corresponding
FIR filter and thus, IIR filter is selected for breathing pattern
extraction. The Butterworth filter is a type of signal processing
filter designed to have as flat frequency response as possible in
the passband. The frequency response of the Butterworth filter is
maximally flat (i.e. has no ripples) in the passband and rolls off
towards zero in the stopband, unlike other filter types that have
non-monotonic ripple in the passband and/or the stopband.
[0055] Thus, the filter is to be designed in such a way that the
noise is efficiently eliminated and the breathing cycle information
is obtained accurately. The coefficients are calculated for these
parameters, and the filtering is done in forward as well as reverse
direction (Zero-phase forward and reverse digital HR filtering),
thereby preserving the frequency information and providing zero
phase distortion. The filter design is perfected over a number of
signals gathered from multiple subjects in relaxed and laboured
breathing scenarios. The cut-off frequencies are decided
empirically because typical breathing rates are much lower like
15-30 times a minute at max as per state-of-the-art. Using the
above mentioned frequencies, the selected filter order, and the
data sampling rate, the system 200 computes the filter coefficients
a & b (in equation 1) using filtfilt function which allows for
a noncausal, zero-phase filtering approach which eliminates the
nonlinear phase distortion of an IIR filter. As is understood, the
filtfilt function performs zero-phase digital filtering by
processing an input data in both the forward and reverse
directions. After filtering the data in the forward direction,
filtfilt reverses the filtered sequence and runs it back through
the filter. The results of filtfilt has characteristics, including,
zero phase distortion, a filter transfer function equal to the
squared magnitude of the original filter transfer function, and a
filter order that is double the order of the filter specified by b
and a.
[0056] The system 200 normalizes the filtered time signal with
respect to its mean and standard deviation (Standard
score=data-mean/stdev) to eliminate amplitude discrepancies which
may lead to differences with the actual breathing patterns. There
are various means of normalizing a time-series data, and the above
technique is one of them. The disclosed normalization technique for
breathing signal is based on an assumption that breathing is a slow
changing bio-marker hence having a zero mean and unit stddev gives
relative amplitude in a better form. Thus, with the help of the
normalized signal, the pre-processed PPG time-signal can be
determined to have high correlation with the actual breathing
pattern signal received from a TBPR device, with the help of their
Pearson Coefficient. The pre-processed PPG signal contains
significant information about the breathing pattern of the subject,
and can be further used for reconstructing breathing cycles and
studying whether the subject has cardiopulmonary related
issues.
[0057] The system 200 analyses the PPG signal, after
pre-processing, in frequency domain, by computing its Fourier
Transform (FFT) and obtaining power spectrum. The number of FFT
points is typically chosen as 512, for a 10 second window and 50 Hz
signal. This can be configured to 2048 points (40 second window)
for getting a better accuracy for windows of interest. From the
power spectrum, the highest isolated peak is taken as breathing
frequency and is used to calculate breathing rate in Breaths per
minute (BrPM). The breathing rate in (BrPM) is the number of
breaths (inhale and exhale cycle) that happen over a time span of 1
minutes.
[0058] Herein the FFT point (i.e. the highest isolated peak) is
selected which allows user to increase resolution for windows of
interest. Thus, the outcome of the disclosed analysis provides both
the average amplitude and frequency of breathing for a 10 second
window of an individual. This can be possibly used for study of
non-specific bio-markers for a number of conditions like sleep
apnea, hypertension, asthma, and so on. An example flow-diagram of
a method for breathing pattern extraction from PPG signal is
described further with reference to FIG. 4.
[0059] FIG. 4 illustrates an example flow diagram of a method 400
for breathing pattern extraction from PPG signal, in accordance
with an example embodiment. Breathing pattern refers to the
amplitude of air flow during a inhale-exhale cycle, as illustrated
in FIG. 3.
[0060] At 402, the method 400 is initiated and raw PPG signal and
accelerometer signals are acquired from a wearable device, for
example a smart watch, at 404. Herein, it will be understood that
for the brevity of discussion, various embodiments of the
disclosure are presented by considering the wearable device as a
smartwatch. However, in various embodiments, the wearable device
may include other such devices that can be worn by the subject and
are capable of embodying a PPG sensor and an accelerometer. The
window size and sampling frequency of the PPG signal are considered
as described with reference to FIG. 2.
[0061] At 406, mean and standard deviation of the accelerometer
signal are computed. Based on the mean and standard deviation of
the accelerometer signal, it is determined at 408, whether the
subject is mobile or not. If it is determined that the subject is
mobile, the PPG signal for said window is discarded at 410, and the
method terminates at 412.
[0062] If however, at 408 it is determined that the subject is not
mobile, the PPG signal is considered for further analysis. For
example, at 414, the PPG signal is interpolated, and its DC
component is filtered at 416 using IIR Butterworth band-pass filter
of order 4, with a cut-off frequency of 0.2-1 Hz. The breathing
cycle is reconstructed at 418 to obtain the breathing pattern, and
breathing rate is computed using the FFT at 420. An example
scenario for describing experimental results corresponding to the
disclosed method and system is explained below.
[0063] FIG. 5 illustrates an example method 500 for breathing
pattern extraction from PPG signal, in accordance with an example
embodiment. The method 500 may be described in the general context
of computer executable instructions. Generally, computer executable
instructions can include routines, programs, objects, components,
data structures, procedures, modules, functions, etc., that perform
particular functions or implement particular abstract data types.
The method 500 may also be practiced in a distributed computing
environment where functions are performed by remote processing
devices that are linked through a communication network. The order
in which the method 500 is described is not intended to be
construed as a limitation, and any number of the described method
blocks can be combined in any order to implement the method 500, or
an alternative method. Furthermore, the method 500 can be
implemented in any suitable hardware, software, firmware, or
combination thereof. In an embodiment, the method 500 depicted in
the flow chart may be executed by a system, for example, the system
200 of FIG. 2. In an example embodiment, the system 200 may be
embodied in an exemplary computer system.
[0064] Referring to FIG. 5, the method for breathing pattern
extraction from PPG signals is initiated at 502, where a filter is
designed for extracting breathing patterns from PPG signals. The
designing of the filter includes defining filter specifications for
extraction of breathing pattern from the PPG signals at 502.
Herein, the filter specifications includes a type, an order and a
cut-off frequency of the filter. Further designing of the filter
includes generating a transfer function associated with the filter
specifications, at 506 and computing a plurality of filter
coefficients using filtfilt function for allowing filtering of the
PPG signals, at 508. In an embodiment the filter is designed by
using the system, for example the system 200 (as is described with
reference to FIG. 2). At 510, the method 500 includes generating
filtered PPG signals using the filter comprising the plurality of
filter coefficients, by removing DC component from PPG signals
obtained from a wearable device being worn by a subject. The
filtered PPG signal are indicative of the breathing pattern of the
subject.
[0065] An example scenario for breathing pattern extraction from
PPG signal is described further in the description by referring to
FIGS. 6-8.
Example Scenario
[0066] In order to validate the accuracy of breathing pattern
extraction method and system, the experimental results are
provided. 38 sessions of data are collected from 19 subjects with
age group of 31.+-.8 years. Two data streams were collected in each
session, one from a TBPS as ground truth, and another from a
Samsung Gear S2.TM. smartwatch which was worn by the subjects on
their wrist during the sessions. The subjects were asked to breathe
at different rates during the session to ensure a wide spectrum of
breathing frequencies is covered. Each session was approximately 60
s of duration. It was analysed and processed the data from 10 s-50
s. It was observed that the method is able to extract the breathing
cycles from PPG very accurately, and shows a very high correlation
to those obtained from the TBP device. A Pearson's Correlation (PC)
is estimated for each session between these two streams of data. PC
is a bivariate tool which expresses the strength of correlation
between two random variable X and Y and is calculated as:
.rho. ( X , Y ) = Cov ( X , Y ) .sigma. X .sigma. Y
##EQU00001##
where Cov is covariance, and a is the standard deviation.
[0067] Following figures, FIGS. 6, 7 and 8 illustrates results are
derived from the experiment:
[0068] FIG. 6 shows PC calculated for each breathing cycle for a
single best session, where the average PC stands at 0.987 proving
the high accuracy of the disclosed breathing cycle extraction
method. It is also observed that the occurrence of respiratory
peaks in the PPG-extracted signal was accurate within an average of
0.5 s, cementing the physiological effect of respiration on the
PPG. FIG. 7 shows PC calculated for each full session, where 30 out
of 38 sessions have an encouraging PC>0.5. The breathing rates
estimated from TBPS and PPG are compared by calculating Mean
Absolute Deviation (MAD) for each session (FIG. 8; 3 outliers
removed). Average MAD for all the sessions was determined to be
only 1.5.
[0069] Various embodiments disclosed herein provide method and
system for breathing pattern extraction from PPG signals. The
disclosed system can be installed in an electronic device such as a
wearable (for example a smart watch), and hence the proposed system
is easily portable and can be utilized for continuous monitoring.
Moreover such wearable devices includes powerful set of sensors,
including IMU and PPG. Hence the user is not subject to any special
hardware. The system uses IMU in order to detect the spells of
intensive activities, and uses PPG to extract the cardiac and
breathing cycle information. The system uses PPG from wrist which
is an indirect marker of Breathing rate. In addition, the system
utilizes a forward-reverse zero phase IIR Butterworth band-pass
filter designed specifically to get the noise-removed PPG signal in
Respiration range.
[0070] In an embodiment, the system involves itself with providing
an opportunistic sensing of the physiological parameters pertaining
to detection of palpitation and dyspnea, and detecting the fatigued
states around the physically intensive spells that the user
undertakes as routine tasks. In addition to a ubiquitous
monitoring, the platform can be used as a longitudinal clinical
assessment tool, where a specialist can direct the subject to
perform certain activities with customisable parameters like
pre-workout resting duration etc. Physiological parameters are
recorded before and after the activity. Additionally information
like subject's current fitness level (normal, moderately sick or
acutely sick), comfort level while doing activity etc. can also be
collected in the form of a questionnaire to assist the specialist
with a more in-depth analytics of the recorded parameters.
[0071] The embodiments herein can comprise hardware and software
elements. The embodiments that are implemented in software include
but are not limited to, firmware, resident software, microcode,
etc. The functions performed by various modules described herein
may be implemented in other modules or combinations of other
modules. For the purposes of this description, a computer-usable or
computer readable medium can be any apparatus that can comprise,
store, communicate, propagate, or transport the program for use by
or in connection with the instruction execution system, apparatus,
or device.
[0072] The medium can be an electronic, magnetic, optical,
electromagnetic, infrared, or semiconductor system (or apparatus or
device) or a propagation medium. Examples of a computer-readable
medium include a semiconductor or solid state memory, magnetic
tape, a removable computer diskette, a random access memory (RAM),
a read-only memory (ROM), a rigid magnetic disk and an optical
disk. Current examples of optical disks include compact disk-read
only memory (CD-ROM), compact disk-read/write (CD-R/W) and DVD.
* * * * *